Building Your First Image Classification -Teachable Machine
Issue 3: Malaysian fruits classification
Welcome back, tech explorers! This week, we'll be diving headfirst into the world of Machine Learning with a project that celebrates the unique flavors of Malaysia! Get ready to build a model that can identify three iconic Malaysian fruits: durian, rambutan, and mangosteen.
What You'll Need:
A computer with internet access
A few minutes of your time
A dash of enthusiasm for AI!
Project Goal: We'll create a model that can distinguish between durian, rambutan, and mangosteen – a fun challenge for anyone who loves exploring the exotic!
Step 1: Data Collection - Gather Your Malaysian Fruit Friends!
The first step in training our model is collecting data. In this case, we'll need images of our chosen Malaysian fruits: durian, rambutan, and mangosteen.
Finding Images: You can use your phone camera to capture pictures of real fruit, or you can search online for royalty-free images. Aim to collect around 10-15 images for each fruit category (durian, rambutan, mangosteen).
Image Quality: Make sure the images are clear and well-lit, with the fruits centered in the frame. It might be helpful to capture images from different angles to give your model a more well-rounded perspective.
Step 2: Labeling - Telling Your Model What's What
Now that you have your images, it's time to tell Teachable Machine what they represent. This is called labeling.
Head over to the Teachable Machine website:
https://teachablemachine.withgoogle.com/
Choose "Image" as your project type.
Click on "Upload" and select the images for your first category (e.g., durian).
In the label box below the image, type "durian".
Repeat steps 3 and 4 for all your durian images.
Do the same for your rambutan and mangosteen images, assigning the labels "rambutan" and "mangosteen" respectively.
Step 3: Training - Teaching Your Model to See Malaysian Fruits
Once you've labeled all your images, it's time for the magic to happen! Teachable Machine will analyze the images and learn to differentiate between the fruits.
Click on the "Train" button.
Teachable Machine will process your data. This might take a few minutes depending on the number of images you uploaded.
Step 4: Testing - Putting Your Model to the Test!
Now that your model is trained, let's see how well it performs!
Click on the "Test" button.
Teachable Machine will display a new image.
Below the image, guess which Malaysian fruit it is (durian, rambutan, or mangosteen).
Click on the "Predict" button.
Teachable Machine will show you its prediction along with a percentage score indicating its confidence level.
Repeat steps 2-5 with several different test images (ideally images not used in the training data) to get a good sense of your model's accuracy.
Interpreting the Results:
The accuracy percentage shown during testing tells you how often your model correctly identified the fruit in the image.
High Accuracy (80% or above): Tahniah (Congratulations)! Your model has learned well and can successfully distinguish between the Malaysian fruits.
Low Accuracy (less than 50%): There might be room for improvement. You can try collecting more images or adjusting training parameters (discussed in Issue 4).
Congratulations! You've just built your first Machine Learning project using Teachable Machine, celebrating the unique fruits of Malaysia. Isn't that exciting?
Bonus Challenge: Can you think of other image classification projects that showcase culture or landmarks? Share your ideas in the comments below!
Challenge:
Imagine you're running a fruit stall at a local market. Could you use your model to create a simple app that helps customers identify different types of Malaysian fruits? Think about how you could display the model's predictions in a user-friendly way.
Poll Time:
Looking Ahead: In the next issue (Issue 4), we'll delve deeper into the world of training! We'll explore advanced options within Teachable Machine that can help you fine-tune your model for even better performance. Stay tuned! Subscribe!
Simple and clear steps..